> That's science. Maybe Noasaurus will prove undiagnostic since it's
> so fragmentary. Better get rid of Noasauridae in favor of
> "Masiakasauridae" (since Velocisaurus is hardly better).
> Abelisaurus is pretty poorly known too. Carnotauridae is the safer
> way to go. Ornithomimus is based on a few metatarsals and
> metacarpals that may not even belong to the same taxon. Let's go
> with Struthiomimidae just to be sure.
You know... all these strike me as being REALLY good ideas. ;-)
For the sake of stability, I really do think it is better to anchor
clades in well-known taxa. And by 'well-known' I mean scientifically
well-known (i.e., not fragmentary).